How much to tighten the H50 cooling head to CPU safely?

I have a ASUS P6X58D Premium X58 Socket MB with a Intel Core i7-960 3.2Ghz 8M LGA1366 CPU.

When tightning down the cooling head of the H50 to the CPU I noticed the four copper feet (supports, whatever) on the locking frame for the cooler did not completely touch the MB on two corners. This was after applying considerable torque to the screws. I was starting to get worried about actually bending (bowing) the MB or crushing the CPU so I stopped with about a 1/16" gap under the H50 locking frame "feet" and the MB surface.

Everything seems okay but I haven't over-clocked the CPU or done any significant stress testing yet. I am not using any 3rd party thermal grease or otherwise (just the white contact patch that was pre-applied to the H50 cooling head).

I wanted to see if anybody else has had difficulty getting the H50 cooling head frame mount to sit flush with the MB on the CPU side and how much torque is enough to SAFELY secure the H50 on the CPU.

Thanks

I have a ASUS P6X58D Premium X58 Socket MB with a Intel Core i7-960 3.2Ghz 8M LGA1366 CPU.

 

When tightning down the cooling head of the H50 to the CPU I noticed the four copper feet (supports, whatever) on the locking frame for the cooler did not completely touch the MB on two corners. This was after applying considerable torque to the screws. I was starting to get worried about actually bending (bowing) the MB or crushing the CPU so I stopped with about a 1/16" gap under the H50 locking frame "feet" and the MB surface.

 

Everything seems okay but I haven't over-clocked the CPU or done any significant stress testing yet. I am not using any 3rd party thermal grease or otherwise (just the white contact patch that was pre-applied to the H50 cooling head).

 

I wanted to see if anybody else has had difficulty getting the H50 cooling head frame mount to sit flush with the MB on the CPU side and how much torque is enough to SAFELY secure the H50 on the CPU.

 

Thanks

All i3/i5/i7 are flip chip, meaning that they are on an array of cantilever pins, you should never exceed Intel's specs for static/dynamic loadings (force and moment), as this might move the contact points from their desired orientation.

Also, applying too much of a moment (tightening one side much more than the other side) will cause some contacts to be overloaded, while others might lose their desired contact points.

Been there, done that. :sigh!:

Check that your H50 hoses are not torquing the heatsink, this usually means orientation of the heatsink to as neutral a position as is possible.

I can post the Intel specs and the maximum amount of torque is based simply on the turning ratio (screwdriver diameter to the screw diameter ratio) times the screw pitch.

More simply put, snug screws should be sufficient assuming the heatsink is plumb with the CPU IHS (which should be true if there are no residual moments between the two).

Too much information?

Not too much info at all and it makes me think that Corsair should put a sticky on this forum or update their H50 install video to address this information.

In the video (my source for install info) it just shows the guy cranking the screws down in no particular order until the H50 mounting frame sat flush with the MB.

Glad I stopped when I did. As a matter of fact, I may go and back the screws off a half turn or so just to be safe.

Just a point of information.

Getting the H50 heatsink and the CPU IHS "flush" can be difficult, as I've learned through several attempts.

You never know for sure until you remove the H50 heatsink. :sigh!: A paradox of sorts.

The method I ended up going with for paste TIMS, was first to screw down the bracket until the four plastic spacers hit the MB. I then backed off a set number of full turns (typically four), enough space to get the H50 heatsink through the bracket. Then I took a thin piece of plastic (like those from a grocery bag) and placed it between the CPU IHS and the H50 copper face, and tightened down the four screws, this simulates the TIM layer. Now I again backed off of the four screws a set number of turns, now knowing where the screws should be once the TIM is applied.

If done right, when you remove the H50 (the paradox) the two faces should actually be stuck together, and it takes some force to twist and pull apart, a suction is created between the two faces if they are parallel with each other and just enough TIM was applied to cover the CPU IHS completely.